CN116403605B - Stacker fault prediction method and related device - Google Patents
Stacker fault prediction method and related device Download PDFInfo
- Publication number
- CN116403605B CN116403605B CN202310673692.5A CN202310673692A CN116403605B CN 116403605 B CN116403605 B CN 116403605B CN 202310673692 A CN202310673692 A CN 202310673692A CN 116403605 B CN116403605 B CN 116403605B
- Authority
- CN
- China
- Prior art keywords
- sound
- sound data
- processed
- stacker
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 99
- 230000007246 mechanism Effects 0.000 claims description 230
- 238000010801 machine learning Methods 0.000 claims description 63
- 230000036541 health Effects 0.000 claims description 61
- 230000008569 process Effects 0.000 claims description 37
- 238000012549 training Methods 0.000 claims description 23
- 230000015654 memory Effects 0.000 claims description 12
- 230000003068 static effect Effects 0.000 claims description 8
- 238000003860 storage Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 5
- 230000036961 partial effect Effects 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 2
- 230000007257 malfunction Effects 0.000 claims 1
- 230000002829 reductive effect Effects 0.000 abstract description 10
- 238000004519 manufacturing process Methods 0.000 abstract description 8
- 238000012423 maintenance Methods 0.000 description 15
- 230000002159 abnormal effect Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 230000000306 recurrent effect Effects 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 7
- 238000003062 neural network model Methods 0.000 description 6
- 230000035945 sensitivity Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 5
- 238000012847 principal component analysis method Methods 0.000 description 5
- 125000004122 cyclic group Chemical group 0.000 description 4
- 238000001914 filtration Methods 0.000 description 4
- 239000000463 material Substances 0.000 description 4
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000000513 principal component analysis Methods 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000000670 limiting effect Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 230000002441 reversible effect Effects 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000008485 antagonism Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Abstract
The application provides an equipment failure prediction method, a stacker failure prediction method and a related device, wherein the equipment failure prediction method comprises the following steps: acquiring sound data to be processed, wherein the sound data to be processed is sound data of the collected target machine equipment in one operation period; predicting target sound characteristics of target machine equipment in the next operation period according to the sound data to be processed; and carrying out fault prediction of the next operation period on the target machine equipment according to the target sound characteristics. According to the method, the target sound characteristics of the target machine equipment in the next operation period are predicted according to the sound data to be processed in the previous operation period, and the fault prediction of the next operation period is carried out on the target machine equipment according to the target sound characteristics, so that the fault possibility can be predicted before the fault of the target machine equipment occurs, the overhaul can be verified manually, and the probability that the normal operation of the production line is not affected in time due to the fact that the fault of the machine equipment is found is effectively reduced.
Description
Technical Field
The application relates to the field of equipment fault prediction, in particular to a stacker fault prediction method and a related device.
Background
In general, a sound is generated when a machine device is operated, and the sound generated when the machine device is failed is different from that generated when the machine device is normal, so that whether the device is a failed device can be found through the sound generated by the machine device. However, if the machine equipment is not maintained in time due to the fact that the abnormal sounds are not found by the operation and maintenance personnel, the equipment has a large fault risk, and once the machine equipment is in fault, the production line where the machine equipment is located cannot operate well.
Disclosure of Invention
The embodiment of the application aims to provide a stacker fault prediction method and a related device, which are used for solving the problem that the normal operation of a production line is not affected in time due to the fact that machine equipment faults are found.
The embodiment of the application provides a device fault prediction method, which comprises the following steps: acquiring sound data to be processed, wherein the sound data to be processed is sound data of the collected target machine equipment in one operation period; predicting target sound characteristics of target machine equipment in the next operation period according to the sound data to be processed; and carrying out fault prediction of the next operation period on the target machine equipment according to the target sound characteristics. In the implementation process of the scheme, the target sound characteristics of the target machine equipment in the next operation period are predicted according to the sound data to be processed in the previous operation period, and the fault prediction of the next operation period is carried out on the target machine equipment according to the target sound characteristics, so that the fault possibility can be predicted before the fault of the target machine equipment occurs, the overhaul can be verified manually, and the probability that the normal operation of the production line is not affected in time due to the fact that the fault of the machine equipment is found is effectively reduced.
Optionally, in an embodiment of the present application, predicting, according to the sound data to be processed, a target sound feature of the target machine device in a next operation period includes: extracting sound characteristics of the sound data to be processed to obtain the sound characteristics to be processed; predicting the sound characteristics to be processed through a machine learning model to obtain target sound characteristics; wherein, the machine learning model is formed according to sample data training, and the sample data includes: the sound characteristics of the sound data of the target machine device in one operation cycle and the sound characteristics of the sound data of the target machine device in the next operation cycle of the operation cycle. In the implementation process of the scheme, the sound characteristics to be processed are predicted through the machine learning model, the target sound characteristics are obtained, the fault prediction of the next operation period is carried out on the target machine equipment according to the target sound characteristics, and the accuracy of predicting the fault of the target machine equipment by utilizing the sound data is improved.
Optionally, in an embodiment of the present application, the type of the target sound feature includes at least one of a time domain feature, a frequency domain feature, and a time frequency feature; if the type of the target sound feature includes a time domain feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including: predicting the to-be-processed sound features of the time domain feature type through a machine learning model to obtain target sound features of the time domain feature type; if the type of the target sound feature includes a frequency domain feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including: predicting the to-be-processed sound features of the frequency domain feature type through a machine learning model to obtain target sound features of the frequency domain feature type; if the type of the target sound feature includes a time-frequency feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including: predicting the to-be-processed sound features of the time-frequency feature type through a machine learning model to obtain target sound features of the time-frequency feature type. In the implementation process of the scheme, the type of the target sound feature comprises at least one of time domain features, frequency domain features and time frequency features, so that the variety of the target sound feature is richer, the prediction result is more accurate through more variety of sound features, and the accuracy of fault prediction is improved.
Optionally, in an embodiment of the present application, predicting, according to the sound data to be processed, a target sound feature of the target machine device in a next operation period includes: dividing the sound data to be processed into the sound data to be processed with different working conditions according to the control signal time sequence data of the target machine equipment; predicting target sound characteristics in the next operation period of each working condition according to the sound data to be processed of the working condition; and aiming at target sound characteristics predicted by the sound data to be processed under a working condition, carrying out fault prediction of a next operation period on a mechanism of the target machine equipment in an operation state under the working condition. In the implementation process of the scheme, the to-be-processed sound data are divided into to-be-processed sound data of different working conditions according to the control signal time sequence data of the target machine equipment, so that a fault mechanism in the next operation period in different working conditions is conveniently positioned, the accuracy of fault prediction is improved, and the efficiency of manually overhauling the fault mechanism is improved.
Optionally, in an embodiment of the present application, according to control signal time sequence data of the target machine device, dividing the to-be-processed sound data into to-be-processed sound data of different working conditions includes: dividing sound data of a first preset working condition of target machine equipment from the sound data to be processed according to the control signal time sequence data; dividing the sound data of the second preset working condition and the sound data of the third preset working condition of the target machine equipment from the sound data except the sound data of the first preset working condition in the sound data to be processed according to the frequency domain characteristic difference of the sound data when all mechanisms of the target machine equipment independently operate; the target machine equipment is different in mechanisms in running states in a first preset working condition, a second preset working condition and a third preset working condition. In the implementation process of the scheme, the voice data of the plurality of preset working conditions of the residual target machine equipment are divided into the voice data of the second preset working condition and the voice data of the third preset working condition according to the frequency domain characteristic difference, so that the granularity of the working condition division is increased.
Optionally, in an embodiment of the present application, the target sound feature includes a plurality of sound features; performing fault prediction of the next operation period on the target machine equipment according to the target sound characteristics, including: and performing fault prediction of the next operation period on the target machine equipment by adopting the sound characteristics sensitive to the faults of the target machine equipment in the plurality of sound characteristics. In the implementation process of the scheme, the fault prediction of the next operation period is carried out on the target machine equipment by adopting the sound characteristics which are sensitive to the fault of the target machine equipment in the plurality of sound characteristics, so that the interference of the sound characteristics which are insensitive to the fault and the operation quantity of the equipment are reduced, and the accuracy rate of the fault prediction of the next operation period on the target machine equipment is effectively improved.
Optionally, in an embodiment of the present application, the sound feature sensitive to the failure of the target machine device is selected from a plurality of sound features by using a principal component analysis method. In the implementation process of the scheme, the sound characteristics sensitive to the faults of the target machine equipment are screened out from the plurality of sound characteristics by adopting the principal component analysis method, so that the interference of the non-principal component characteristics in the plurality of sound characteristics is reduced, and the accuracy of fault prediction of the next operation period of the target machine equipment is effectively improved.
Optionally, in an embodiment of the present application, performing fault prediction of the next operation cycle on the target machine device according to the target sound feature includes: determining a health index of the target machine device according to the target sound characteristics; and carrying out fault prediction of the next operation period on the target machine equipment according to the health index. In the implementation process of the scheme, the fault prediction of the next operation period is carried out on the target machine equipment according to the health index of the target machine equipment determined by the sound characteristics, so that the fault of the target machine equipment is effectively quantized into the health index, and the accuracy of the fault prediction is improved. In some scenes, the health management can be performed on the equipment according to the health index, so that the maintenance and management capacity of the equipment is improved.
Optionally, in an embodiment of the present application, after determining the health index of the target machine device according to the target sound feature, the method further includes: and if the health index is in the alarm value interval, generating alarm information. In the implementation process of the scheme, the alarm information is generated under the condition that the health index is located in the alarm value interval, so that the condition that the fault alarm is difficult when the abnormal sound is small is improved, and the fault probability of the target machine equipment is effectively improved.
The embodiment of the application also provides a stacker fault prediction method, which comprises the following steps: acquiring sound data to be processed, wherein the sound data to be processed are sound data of a horizontal running mechanism, a lifting running mechanism and a fork telescopic mechanism in an operation period in the acquired stacker; predicting target sound characteristics of the stacker in the next operation period according to the sound data to be processed; and carrying out fault prediction on the mechanism of the stacker in the next operation period according to the target sound characteristics. In the implementation process of the scheme, the target sound characteristic of the stacker in the next operation period is predicted according to the sound data to be processed in the previous operation period, and the fault prediction of the stacker in the next operation period is performed according to the target sound characteristic, so that the fault possibility can be predicted before the stacker faults occur, the overhaul can be verified manually, and the probability that the normal operation of the production line is not affected in time due to the fact that the stacker faults are found is effectively reduced.
Optionally, in an embodiment of the present application, predicting, according to the sound data to be processed, a target sound characteristic of the stacker in a next operation cycle includes: extracting sound characteristics of the sound data to be processed to obtain the sound characteristics to be processed; predicting the sound characteristics to be processed through a machine learning model to obtain target sound characteristics; wherein, the machine learning model is formed according to sample data training, and the sample data includes: the sound characteristics of sound data of the stacker in one operation cycle and the sound characteristics of sound data of the stacker in the next operation cycle of the operation cycle. In the implementation process of the scheme, the sound characteristics to be processed are predicted through the machine learning model, the target sound characteristics are obtained, the fault prediction of the next operation period is carried out on the stacker according to the target sound characteristics of the next operation period, and the accuracy of predicting the fault of the stacker by utilizing sound data is improved.
Optionally, in an embodiment of the present application, the type of the target sound feature includes at least one of a time domain feature, a frequency domain feature, and a time frequency feature; if the type of the target sound feature includes a time domain feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including: predicting the to-be-processed sound features of the time domain feature type through a machine learning model to obtain target sound features of the time domain feature type; if the type of the target sound feature includes a frequency domain feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including: predicting the to-be-processed sound features of the frequency domain feature type through a machine learning model to obtain target sound features of the frequency domain feature type; if the type of the target sound feature includes a time-frequency feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including: predicting the to-be-processed sound features of the time-frequency feature type through a machine learning model to obtain target sound features of the time-frequency feature type.
Optionally, in an embodiment of the present application, predicting, according to the sound data to be processed, a target sound characteristic of the stacker in a next operation cycle includes: dividing the sound data to be processed into the sound data to be processed with different working conditions according to the control signal time sequence data of the stacker; predicting target sound characteristics in the next operation period of each working condition according to the sound data to be processed of the working condition; wherein, different operating modes include: the horizontal running mechanism running condition, the lifting running mechanism running condition and the fork telescopic mechanism running condition; aiming at the target sound characteristics predicted by the sound data to be processed of a working condition, carrying out fault prediction of the next operation period on a mechanism of the stacker in an operation state in the working condition. In the implementation process of the scheme, the to-be-processed sound data are divided into the sound data of different working conditions according to the control signal time sequence data of the stacker, so that a fault mechanism of different working conditions in the next operation period is conveniently positioned, and the efficiency of manually overhauling the fault mechanism is improved.
Optionally, in an embodiment of the present application, dividing the to-be-processed sound data into to-be-processed sound data of different working conditions according to control signal time sequence data of the stacker includes: searching acquisition time meeting a first preset condition from control signal time sequence data of the stacker, wherein the first preset condition comprises the following steps: the fork telescopic mechanism is in a motion state in the operation period, and the horizontal operation mechanism and the lifting operation mechanism are in a static state in the operation period; searching sound data meeting the first preset condition at the acquisition time from the sound data to be processed; and dividing the sound data meeting the first preset condition at the acquisition time into sound data of the operation working condition of the fork telescopic mechanism. In the implementation process of the scheme, the sound data meeting the first preset condition at the acquisition time is divided into the sound data of the operating condition of the fork telescopic mechanism, so that the granularity of the operating condition division is increased.
Optionally, in an embodiment of the present application, dividing the to-be-processed sound data into to-be-processed sound data of different working conditions of the stacker according to control signal time sequence data of the stacker includes: searching acquisition time meeting a second preset condition from control signal time sequence data of the stacker, wherein the second preset condition comprises the following steps: the fork telescopic mechanism is in a static state in the running period, and the horizontal running mechanism and the lifting running mechanism are in a moving state in the running period; searching sound data meeting the second preset condition at the acquisition time from the sound data to be processed; according to the frequency domain characteristic difference of the sound data when each mechanism of the stacker operates independently, the sound data meeting the second preset condition at the acquisition time is divided into the sound data of the operating condition of the horizontal operating mechanism and the sound data of the operating condition of the lifting operating mechanism.
Optionally, in an embodiment of the present application, the target sound feature includes a plurality of sound features; and carrying out fault prediction on a mechanism of the stacker for the next operation period according to the target sound characteristics, wherein the fault prediction comprises the following steps: and respectively carrying out fault prediction on each mechanism of the stacker by adopting the sound characteristics which are sensitive to faults of each mechanism of the stacker in the plurality of sound characteristics. In the implementation process of the scheme, the fault prediction of each mechanism of the stacker is respectively carried out on each mechanism of the stacker by adopting the sound characteristics sensitive to the faults of each mechanism of the stacker in the plurality of sound characteristics, so that the sound characteristics insensitive to the faults and the operation amount of equipment are reduced, and the accuracy of the fault prediction of the next operation period of the target machine equipment is effectively improved.
Optionally, in an embodiment of the present application, the acoustic features that are sensitive to faults of the various mechanisms of the stacker are screened using principal component analysis. In the implementation process of the scheme, the sound characteristics sensitive to the faults of the target machine equipment are screened out from the plurality of sound characteristics by adopting the principal component analysis method, so that the interference of the non-principal component characteristics in the plurality of sound characteristics is reduced, and the accuracy of fault prediction of the next operation period of the target machine equipment is effectively improved.
Optionally, in an embodiment of the present application, performing fault prediction on a mechanism of the stacker for a next operation period according to the target sound feature includes: determining health indexes of all institutions of the stacker according to the target sound characteristics; and carrying out fault prediction on each mechanism of the stacker for the next operation period according to the health index. In the implementation process of the scheme, the fault prediction of each mechanism of the stacker is carried out for the next operation period according to the health index of the target machine equipment determined according to the sound characteristics, so that the fault of the stacker is effectively quantized into the health index, and the accuracy of the fault prediction is improved.
Optionally, in an embodiment of the present application, after determining the health index of each organization of the stacker according to the target sound feature, the method further includes: and if the health index is in the alarm value interval, generating alarm information. In the implementation process of the scheme, the alarm is carried out under the condition that the health index is located in the alarm value interval, so that the condition that the fault alarm is difficult when the abnormal sound is small is improved, and the fault probability of the stacker is effectively improved.
The embodiment of the application also provides a device fault prediction system, which comprises: a sound sensor and a computing device; the sound sensor is used for collecting sound data to be processed, wherein the sound data to be processed is the sound data of the target machine equipment in one operation period; the computing device is used for predicting target sound characteristics of the target machine device in the next operation period according to the sound data to be processed; and carrying out fault prediction of the next operation period on the target machine equipment according to the target sound characteristics.
Optionally, in an embodiment of the present application, the computing device includes an edge device and a server; the edge device is used for: extracting sound characteristics of the sound data to be processed to obtain the sound characteristics to be processed; the server is used for: predicting the sound characteristics to be processed through a machine learning model to obtain target sound characteristics; performing fault prediction of the next operation period on the target machine equipment according to the target sound characteristics; wherein, the machine learning model is formed according to sample data training, and the sample data includes: the sound characteristics of the sound data of the target machine device in one operation cycle and the sound characteristics of the sound data of the target machine device in the next operation cycle of the operation cycle. In the implementation process of the scheme, the edge equipment extracts the sound characteristics of the sound data to be processed, so that the calculation pressure and load of the server are reduced.
Optionally, in the embodiment of the present application, the edge device is specifically configured to: dividing the sound data to be processed into the sound data to be processed with different working conditions according to the control signal time sequence data of the target machine equipment; extracting sound characteristics of the sound data to be processed in each working condition to obtain the sound characteristics to be processed; the method comprises the steps of predicting a target sound characteristic predicted by sound data to be processed under a working condition, and predicting a fault of a next operation period of a mechanism of target machine equipment in an operation state under the working condition.
Optionally, in the embodiment of the present application, the edge device is specifically configured to: dividing sound data of the operating condition of an independent mechanism of the target machine equipment from the sound data to be processed according to the control signal time sequence data; according to the frequency domain characteristic difference of the sound data when each mechanism of the target machine equipment operates independently, the sound data of the working condition of the multi-mechanism operation of the residual target machine equipment is divided into the to-be-processed sound data of the working condition when each mechanism operates independently.
Optionally, in the embodiment of the present application, the edge device is specifically configured to: dividing sound data of a first preset working condition of target machine equipment from the sound data to be processed according to the control signal time sequence data; dividing the sound data of the second preset working condition and the sound data of the third preset working condition of the target machine equipment from the sound data except the sound data of the first preset working condition in the sound data to be processed according to the frequency domain characteristic difference of the sound data when all mechanisms of the target machine equipment independently operate; the target machine equipment is different in mechanisms in running states in a first preset working condition, a second preset working condition and a third preset working condition.
Optionally, in an embodiment of the present application, the target sound feature includes a plurality of sound features; the server is specifically configured to: and performing fault prediction of the next operation period on the target machine equipment by adopting the sound characteristics sensitive to the faults of the target machine equipment in the plurality of sound characteristics.
Optionally, in an embodiment of the present application, performing fault prediction of the next operation cycle on the target machine device according to the target sound feature includes: determining a health index of the target machine device according to the target sound characteristics; and carrying out fault prediction of the next operation period on the target machine equipment according to the health index.
Optionally, in an embodiment of the present application, the server is further configured to: and if the health index is in the alarm value interval, generating alarm information.
The embodiment of the application also provides a stacker, which comprises: the device comprises a horizontal running mechanism, a lifting running mechanism, a fork telescopic mechanism and a sound sensor, wherein the fork telescopic mechanism is arranged on the lifting running mechanism; the sound sensor is arranged at least one mechanism of the horizontal running mechanism, the lifting running mechanism and the fork telescopic mechanism and is used for collecting sound data when the stacker runs.
Optionally, in an embodiment of the present application, the sound sensor includes a first part of the sound sensor disposed on the horizontal running mechanism; the first partial sound sensor is arranged at the position of a part to be monitored of the adjacent horizontal running mechanism, and the part to be monitored comprises: at least one of a horizontal motor, a gear rack, a top rail and a bottom rail.
Optionally, in an embodiment of the present application, the sound sensor includes a second part of the sound sensor disposed on the lifting operation mechanism; the second partial sound sensor is arranged at the position of a part to be monitored of the adjacent lifting operation mechanism, and the part to be monitored comprises: at least one of a lift motor and a sprocket chain.
Optionally, in an embodiment of the present application, the sound sensor includes a third part of the sound sensor disposed on the fork telescopic mechanism; the third part sound sensor sets up in the position of the part of waiting to monitor of close fork telescopic machanism, waits to monitor the part and includes: at least one of a telescoping motor and a telescoping plate.
Embodiments of the present application also provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs a method as described above.
Additional features and advantages of embodiments of the application will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of embodiments of the application.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments of the present application will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application, and therefore should not be considered as limiting the scope, and other related drawings can be obtained according to these drawings without inventive effort to those of ordinary skill in the art.
Fig. 1 is a schematic flow chart of an equipment failure prediction method according to an embodiment of the present application;
FIG. 2 is a schematic diagram of control signal timing data according to an embodiment of the present application;
Fig. 3 is a schematic flow chart of a stacker fault prediction method according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a stacker provided in an embodiment of the present application;
FIG. 5 illustrates a schematic diagram of a computing device provided by an embodiment of the present application.
Icon: 300-stacker; 310-horizontal running mechanism; 311-day rails; 312-rack and pinion; 313-ground rail; 320-lifting operation mechanism; 321-lifting motor; 322-sprocket chain; 330-fork telescoping mechanism; 331-telescoping plate; 332-a telescopic motor; 400-computing device; 410-a processor; 420-memory; 430-storage medium.
Detailed Description
Embodiments of the technical scheme of the present application will be described in detail below with reference to the accompanying drawings. The following examples are only for more clearly illustrating the technical aspects of the present application, and thus are merely examples, and are not intended to limit the scope of the present application.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having" and any variations thereof in the description of the application and in the description of the drawings above are intended to cover non-exclusive inclusions.
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it should be understood that the accompanying drawings in the embodiments of the present application are only for the purpose of illustration and description, and are not intended to limit the scope of the embodiments of the present application. In addition, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in the embodiments of the present application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flow diagrams may be implemented out of order and that steps without logical context may be performed in reverse order or concurrently. Moreover, one or more other operations may be added to or removed from the flow diagrams by those skilled in the art under the direction of the teachings of the embodiments of the present application.
In addition, the described embodiments are only some, but not all, of the embodiments of the present application. The components of the embodiments of the present application generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Accordingly, the following detailed description of the embodiments of the application, as presented in the figures, is not intended to limit the scope of the claimed embodiments of the application, but is merely representative of selected embodiments of the application.
In the description of embodiments of the present application, the technical terms "first," "second," and the like are used merely to distinguish between different objects and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated, a particular order or a primary or secondary relationship. In the description of the embodiments of the present application, the meaning of "plurality" is two or more unless explicitly defined otherwise. In the description of the embodiments of the present application, the term "and/or" is merely an association relationship describing an association object, and indicates that three relationships may exist, for example, a and/or B may indicate: a exists alone, A and B exist together, and B exists alone. In addition, the character "/" herein generally indicates that the front and rear associated objects are an "or" relationship. The term "plurality" refers to two or more (including two), and similarly, "plurality" refers to two or more (including two).
In the related art, an equipment maintainer can find out whether an equipment is malfunctioning through a sound made by a machine equipment. However, when the equipment is found out to be faulty when the equipment gives out fault sounds, maintenance and replacement are carried out, and the production line operation is affected.
For example, in a power exchange station, a stacker is used for storing and transporting batteries, when the stacker operates, components such as a motor, gears, a chain, a rail and the like can generate sounds, when the components are damaged or are bad, the generated sounds of the stacker are abnormal, and an experienced operation and maintenance person can find out equipment faults when hearing the abnormal sounds. However, if the equipment is not maintained in time due to the fact that the abnormal sounds are not found by the operation and maintenance personnel, the equipment has a large fault risk, and once the equipment is in fault, the power exchange station cannot operate well. Moreover, the operation and maintenance personnel can only detect when the device has obvious abnormal sound, and the device can be in bad operation for a long time, the service life of the device can be influenced by long-term bad operation, and the quality of products operated by the device can be influenced by some scenes.
In view of the above problems, in the embodiment of the present application, the target sound data characteristics of the target machine equipment in the next operation period are predicted by the sound data of the target machine equipment in the one operation period, so that the equipment failure is predicted according to the predicted target sound data characteristics, the failure possibility of the equipment in the next operation period can be predicted more timely, and the probability that the normal operation of the production line is affected due to the failure of the target machine equipment is reduced. In some scenes, the fault risk can be found in the early stage of the bad operation of the target machine equipment, the bad operation can be checked in time, and the service life of the target machine equipment is prolonged.
It should be noted that the target machine device in the embodiment of the present application is not limited to the stacker illustrated in the foregoing, but may be other devices that emit sound during operation, such as devices including components such as a motor, a rack and pinion, a chain, a track, and the like, for example, a rotary platform, a die-cutting machine, a winder, and the like.
Please refer to a flow chart of an apparatus fault prediction method provided by an embodiment of the present application shown in fig. 1; the method may be performed by a computing device, which herein refers to an edge device (e.g., an edge console) and/or server having the functionality to perform a computer program, and may include three cases: in the first case, all steps of the method are performed by the edge device; in the second case, all steps of the method are performed by the server; in a third case, part of the steps of the method are performed by the edge device and another part of the steps of the method are performed by the server. Wherein the edge device is for example: data acquisition box, smart phone, personal computer, tablet computer, personal digital assistant, edge industrial control terminal or mobile internet equipment, etc. A server refers to a device that provides computing services over a network, such as: an x86 server and a non-x 86 server, the non-x 86 server comprising: mainframe, minicomputer, and UNIX servers.
The equipment fault prediction method comprises the following steps:
Step S110: and acquiring sound data to be processed, wherein the sound data to be processed is the sound data of the collected target machine equipment in one operation period.
The to-be-processed sound data can be collected by the sound sensor, and the computing device executing the device fault prediction method can directly obtain the to-be-processed sound data from the sound sensor, or can obtain the to-be-processed sound data through the transit device or the storage device.
The sound sensor may be provided on the target machine device or in an environment in which the target machine device operates, for example, within a working range of the target machine device. Optionally, the sound sensor is arranged at a location suitable for capturing sound generated in the operational state of the target machine device, e.g. in the vicinity of a component of the target machine device from which the main sound originates.
It should be understood that one sound sensor may be provided to collect sound data for one target machine device, and a plurality of sound sensors may be provided to collect sound data. In the latter case, the positions of the different sound sensor settings may be different in order to acquire sound data of the device to be processed in all directions by means of the plurality of sound sensors. Or a plurality of sound sensors are respectively arranged for different parts of the target machine equipment for generating sound, so that the sound generated by different parts can be more accurately collected.
An operating cycle may refer to a period of time that the target machine device is operating, e.g., one operating cycle per one hour of target machine device operation; an execution cycle may also be a cycle in which the target machine device performs a workflow, such as: one process of picking and stocking by the stacker is one operation cycle. Of course, one run cycle may also be a cycle in which the target machine device executes a plurality of workflows in succession.
Step S120: and predicting target sound characteristics of the target machine equipment in the next operation period according to the sound data to be processed.
Sound features, which are features extracted from sound data, include, but are not limited to: time domain features, frequency domain features, and/or time-frequency features, etc., where time-frequency features include, but are not limited to: instantaneous frequency, amplitude characteristics, and entropy characteristics, etc.
The next operation cycle refers to the next operation cycle of the collected operation cycle of the sound data to be processed, for example: the target machine device performs a first inventory, a first pick, a second inventory, a second pick, the first inventory and the first pick may be one run, and the second inventory and the second pick may be the next run.
The manner of predicting the target sound characteristic of the target machine device in the next operation period according to the sound data to be processed may be:
in mode 1, target sound data of the next operation cycle is predicted from sound data to be processed by a prediction algorithm, and then target sound features are extracted from the target sound data.
And 2, predicting target sound characteristics of the next operation period according to the sound characteristics to be processed by adopting a prediction algorithm from the sound characteristics to be processed of the sound data to be processed.
It should be appreciated that the prediction algorithms employed in the different ways described above are different.
Step S130: and carrying out fault prediction of the next operation period on the target machine equipment according to the target sound characteristics.
Since the sound characteristics of the faulty device and the non-faulty device are different, the sound characteristics of the well-functioning device and the poorly-functioning device are also different, so that the possibility of the fault of the target machine device in the next operation period can be predicted according to the target sound characteristics of the next operation period.
It should be noted that, each step in the above device fault prediction method may be performed by the same computing device, for example: the steps in the device failure prediction method are all executed by the edge device receiving the sound data to be processed, or may be all executed by the server receiving the sound data to be processed. Similarly, the steps in the above method for predicting an equipment failure may be performed by different equipment, for example: the above-described step S110 and step S120 are performed by an edge device receiving sound data to be processed, and the above-described step S130 is performed by a server extracting sound features of the sound data. Or the above step S110 is performed by an edge device, and steps S120 and S130 are performed by a server. Therefore, the execution subject of each step in the above-described equipment failure prediction method should not be construed as a limitation of the embodiment of the present application.
In the implementation process of the scheme, the target sound characteristics of the target machine equipment in the next operation period are predicted according to the sound data to be processed in the previous operation period, and the fault prediction of the next operation period is carried out on the target machine equipment according to the target sound characteristics, so that the fault possibility can be predicted before the fault of the target machine equipment occurs, the overhaul can be verified manually, and the probability that the normal operation of the production line is not affected in time due to the fact that the fault of the machine equipment is found is effectively reduced.
As an alternative embodiment of the above step S120, an embodiment of predicting the target sound characteristic of the target machine device in the next operation period according to the sound data to be processed may include:
step S121: and extracting sound characteristics of the sound data to be processed to obtain the sound characteristics to be processed.
It can be understood that after the to-be-processed sound data is obtained, filtering and noise reduction can be performed on the to-be-processed sound data to obtain the sound data after filtering and noise reduction. And then, extracting sound characteristics of the sound data after filtering and noise reduction to obtain the sound characteristics to be processed.
Step S122: predicting the sound characteristics to be processed through a machine learning model to obtain target sound characteristics.
Wherein the machine learning model may be trained using sample data, where the sample data may include: the sound characteristics of the sound data of the target machine device in one operation cycle and the sound characteristics of the sound data of the target machine device in the next operation cycle of the operation cycle.
The sound data corresponding to the sample data may be collected for the same target machine equipment, or may be collected for a plurality of similar target machine equipment. Optionally, in the latter case, the specifications of the sound sensors for collecting different target machine devices are the same or similar, and the setting positions of the sound sensors are the same or similar.
Optionally, the specification of the sound sensor for collecting the sound data to be processed is the same as or similar to the specification of the sound sensor for collecting the sample data, and the setting position of the sound sensor for collecting the sound data to be processed is the same as or similar to the setting position of the sound sensor for collecting the sample data.
The machine learning model may employ a recurrent neural network (Recurrent Neural Network, RNN), where the RNN specifically includes: long Short-Term Memory (LSTM) network models or two-way Long-Term Memory (Bidirectional Long Short-Term Memory, bi-LSTM) network models, and the like. Other models suitable for prediction in the acoustic field may also be used as machine learning models.
In the training process of the machine learning model, the sample data (may be sound data including the full life cycle of the stacker) may be split into a training set of a first preset ratio (e.g., 70%), a verification set of a second preset ratio (e.g., 20%), and a test set of a third preset ratio (e.g., 10%), and the training set, the verification set, and the test set may be used to train the machine learning model.
Taking a stacker as an example, a process of acquiring sample data will be described. The full life cycle of the stacker can be a full life cycle of a stacker acceleration experiment simulated in a laboratory, the experiment can be a stacker battery access acceleration experiment, and the experiment aims to quickly reach the end of the stacker life through simulating normal working conditions. The full life cycle may be two charge bit picking and placing cycles, or may be multiple charge bit picking and placing cycles, where two charge bit picking and placing cycles are described as an example, and the stacker performs the steps of the cycle, for example: the battery is removed from the first charge of the battery rack, the battery is placed on the first RGV cart, then the battery is removed from the RGV cart and placed on the second charge, then the battery is removed from the second charge and placed on the second RGV cart, and finally the battery is removed from the second RGV cart and placed on the first charge of the battery rack. And collecting the sound emitted by the stacker in the full life cycle process through the sound sensor, so that sample data can be obtained.
It should be appreciated that the sample data need not be the sound characteristics of the full life cycle sound data described above, but may be sound characteristics of sound data of a portion of the operating cycle during operation of the target machine device.
The training process of machine learning is described below using a recurrent neural network as an example. And calculating the sound characteristics of the sound data in the previous operation period by using the cyclic neural network to obtain calculated sound characteristics, and updating the network weight parameters of the cyclic neural network according to the loss value between the calculated sound characteristics and the sound characteristics of the sound data in the next operation period until the accuracy of the cyclic neural network is not increased or the number of iterations (epoch) is greater than a preset threshold value, thereby obtaining the trained cyclic neural network model. The preset threshold may be set according to the specific situation, for example, set to 100 or 1000, etc. Optionally, in the process of training the recurrent neural network model, the super-parameters of the recurrent neural network model can be adjusted, for example, the optimal combination of super-parameters such as the learning rate is adjusted, so as to shorten the training time for training the recurrent neural network model.
In the training process of the machine learning model (for example, training a neural network model), working condition division may also be performed first, for example, the sound data may be divided into sound data in a horizontal movement working condition, a lifting movement working condition, a telescopic movement working condition and the like according to a combination of frequency domains, frequencies and/or amplitudes of the sound data. Since the acoustic waveforms generated by the vibration of different objects have unique frequency and amplitude combinations, the objects can be identified and classified by analyzing the frequency domain characteristics of the acoustic signals. By virtue of this feature, it is possible to distinguish sound data emitted by two or more mechanisms that move simultaneously by virtue of differences in frequency domain characteristics of sound data when the different mechanisms are operated individually.
Optionally, after the working conditions are divided, the data samples can be classified according to the working conditions, and the samples of each type of working conditions are used for training a machine learning model corresponding to the samples, namely, the sample data of each type of working conditions is used for training a machine learning model for predicting the target sound characteristics of the working conditions. Of course, sample data of different types of working conditions can be used for training a unified machine learning model, and target sound characteristics of different types of working conditions can be predicted through the unified trained machine learning model.
In some embodiments, it is possible that some sample data has no sample tag, i.e., some sample data has only the sound characteristics of sound data in one run period, and no sound characteristics of sound data in the next run period. The recurrent neural Network may be trained in a weakly supervised learning manner at this time, for example, using a heuristics-based or Generation Antagonism Network (GAN) model to generate sample labels (e.g., acoustic features of acoustic data in the next operation period) from current sample data (e.g., acoustic features of acoustic data in the operation period). Of course, a data set with sample tags may also be constructed based on heuristic rules, using an unsupervised clustering algorithm, using an external knowledge base, etc.
Optionally, during the training process of the machine learning model, sound features sensitive to the faults of the target machine equipment can be screened. For example: the method comprises the steps of screening out sound features with fault sensitivity exceeding a threshold value from a plurality of sound features by adopting a Principal Component Analysis (PCA) method, wherein the method can be understood as that the fault sensitivity value of each sound feature in the plurality of sound features is calculated, and then judging whether the fault sensitivity value of the sound feature is larger than a preset threshold value. And if the fault sensitivity level value of the sound feature is larger than a preset threshold value, determining the sound feature as the sound feature sensitive to the fault of the target machine equipment. Similarly, if the fault sensitivity value of the sound feature is less than or equal to a preset threshold, the sound feature is determined to be a sound feature that is not sensitive enough to the fault of the target machine device, where the preset threshold may be set according to an empirical value or may be determined through experiments.
Optionally, the types of the target sound features may include at least one of time domain features, frequency domain features, and time frequency features. When the target sound characteristics are predicted through the machine learning model, one type of sound characteristics to be processed is input, and the same type of target sound characteristics can be obtained.
For example: and if the type of the target sound feature comprises the time domain feature, predicting the to-be-processed sound feature of the time domain feature type through a machine learning model to obtain the target sound feature of the time domain feature type.
If the type of the target sound feature comprises the frequency domain feature, predicting the to-be-processed sound feature of the frequency domain feature type through a machine learning model to obtain the target sound feature of the frequency domain feature type.
If the type of the target sound feature comprises the time-frequency feature, predicting the to-be-processed sound feature of the time-frequency feature type through a machine learning model to obtain the target sound feature of the time-frequency feature type.
As an alternative embodiment of the above step S120, an embodiment of predicting the target sound characteristic of the target machine device in the next operation period according to the sound data to be processed may include:
Step S123: and dividing the sound data to be processed into the sound data to be processed with different working conditions according to the control signal time sequence data of the target machine equipment.
The control signal time sequence data is time sequence data of the control signal, and working conditions of the motion state of the target machine equipment, such as a horizontal motion working condition, a lifting motion working condition, a telescopic motion working condition, a rotary motion working condition and the like, can be determined according to the control signal. Therefore, according to the matching of the time information of the time sequence data of the control signal and the acquisition time of the sound data to be processed, the sound data of which working condition is one time period of the sound data to be processed can be determined, and the sound data to be processed can be further divided into the sound data to be processed of different working conditions.
Step S124: and predicting target sound characteristics in the next operation period of each working condition according to the sound data to be processed of the working condition.
It can be understood that the above-mentioned to-be-processed sound data of each working condition may directly use the same unified training machine learning model to predict the target sound feature in the next operation period of the working condition, and of course, in a specific practical process, different respective training machine learning models may also be used to predict the target sound feature in the next operation period of the working condition.
And aiming at target sound characteristics predicted by the sound data to be processed under a working condition, carrying out fault prediction of a next operation period on a mechanism of the target machine equipment in an operation state under the working condition. In the implementation process of the scheme, the to-be-processed sound data are divided into to-be-processed sound data with different working conditions according to the control signal time sequence data of the target machine equipment, so that a fault mechanism of different working conditions in the next operation period is conveniently positioned, and the efficiency of manually overhauling the fault mechanism is improved.
As an alternative embodiment of the step S123, the embodiment of dividing the to-be-processed sound data into to-be-processed sound data of different working conditions according to the control signal time sequence data of the target machine device may include:
step S123a: and dividing the sound data of the first preset working condition of the target machine equipment from the sound data to be processed according to the control signal time sequence data.
Step S123b: according to the frequency domain characteristic difference of the sound data when each mechanism of the target machine equipment independently operates, the sound data of the second preset working condition and the sound data of the third preset working condition of the target machine equipment are divided from the sound data except the sound data of the first preset working condition in the sound data to be processed.
The target machine equipment is different in mechanisms in running states in the first preset working condition, the second preset working condition and the third preset working condition.
Taking the control signal time sequence data of the stacker shown in fig. 2 as an example; the control signal timing data of the stacker may include: the period time of the stacker in the operation period and the positions of all mechanisms (such as a horizontal operation mechanism of an X axis, a lifting operation mechanism of a Z axis and a fork telescopic mechanism of a fork shaft) of the stacker are determined, and the operation state corresponding to the preset period time of each mechanism of the stacker in the operation period is determined. In the 0 th to 10 th seconds of the time sequence data of the control signals, the positions of the stacker on the X axis, the Z axis and the fork axis do not move, and the sound data collected in the time range of the stacker are not sound data of any one of the three working conditions (horizontal operation mechanism operation working condition, lifting operation mechanism operation working condition and fork telescopic mechanism operation working condition). Similarly, for example: in the 10 th to 15 th seconds of the time sequence data of the control signal, the stacker does not move in the X-axis and Z-axis positions, but moves in the position on the fork shaft, so that the sound data collected by the stacker in the time range is the sound data belonging to the operating condition of the fork telescopic mechanism, and the operating condition of the fork telescopic mechanism can be understood as the first preset operating condition. Because the components (such as a gear rack) in the horizontal running mechanism and the components (such as a chain wheel and a chain) in the lifting running mechanism are different in material, the frequency domain characteristics of sound emitted by the two components are different, so that sound data emitted by the horizontal running mechanism in X-axis motion (namely the running working condition of the horizontal running mechanism) and sound data emitted by the lifting running mechanism in Z-axis motion (namely the running working condition of the lifting running mechanism) can be distinguished according to the frequency domain characteristic differences of the two components; the horizontal operating mode may be understood as the second preset operating mode, and the lifting operating mode may be understood as the third preset operating mode.
Since the acoustic waveforms generated by the vibration of different objects have unique frequency and amplitude combinations, the objects can be identified and classified by analyzing the time domain features and/or frequency domain features of the acoustic signals. By means of this feature, the sound data emitted by two or more mechanisms moving simultaneously can be distinguished by means of differences in the time-domain and/or frequency-domain characteristics of the sound data when the different mechanisms are operated separately.
In the implementation process of the scheme, the sound data of the working conditions of the multi-mechanism operation of the target machine equipment are divided into the to-be-processed sound data of the working conditions of the independent operation of each mechanism according to the frequency domain characteristic difference, so that the granularity of the working condition division is increased.
As an optional implementation manner of the device fault prediction method, the target sound feature includes at least one of a time domain feature, a frequency domain feature and a time frequency feature.
The time domain features may be root mean square values, crest factors, kurtosis, and/or the like.
The frequency domain features may be extracted by: the method comprises the steps of carrying out Fourier transformation on sound data to obtain frequency domain characteristics, wherein the frequency domain characteristics refer to abstract characteristics of a certain signal in a frequency domain, and the abstract characteristics are usually obtained by extracting, calculating and analyzing characteristic parameters of the sound data in the frequency domain through a plurality of specific analysis algorithms. This embodiment specifically includes, for example: the sound data is fourier transformed to obtain frequency domain characteristics such as center of gravity frequency (Centroid Frequency), frequency domain Amplitude average value (Mean Amplitude), and/or frequency variance (Frequency Variance). The frequency domain features described above may also be frequency bands, spectral densities, peaks, powers etc. of the sound data signal, characteristic parameters describing the frequency domain features and properties of the signal.
The extraction mode of the time-frequency characteristics can be as follows: decomposing and transforming the sound data to obtain time-frequency characteristics; wherein, the time-frequency characteristic may include: instantaneous frequency, amplitude characteristics, and entropy characteristics. The specific extraction mode of the time-frequency characteristic is as follows: the sound data is subjected to modal decomposition by using an empirical mode decomposition (EMPIRICAL MODE DECOMPOSITION, EMD) method, so that an original signal of the sound data is decomposed into a plurality of connotation modal components (INTRINSIC MODE FUNCTIONS, IMF), and Hilbert transformation can be further performed on the plurality of connotation modal components (IMF) to obtain instantaneous frequency, amplitude characteristics and (EMD) entropy value characteristics of the working condition category.
In the implementation process of the scheme, the instantaneous frequency, the amplitude characteristic and the entropy characteristic of the working condition category are obtained by carrying out Hilbert transformation on a plurality of connotation modal components, so that the variety of the time-frequency characteristic is richer, the prediction result is more accurate through more kinds of sound characteristics, and the accuracy of fault prediction is improved.
Optionally, after obtaining the above-mentioned time domain features, frequency domain features, and/or time-frequency features, the various sound features may also be stored in a database or imported into other information management systems (e.g., health management platform systems) so that the sound data in the database or information management platform may be used in updating iterations of the machine learning model (e.g., recurrent neural network model).
As an alternative embodiment of the step S130, the target sound feature includes a plurality of sound features; embodiments of performing fault prediction for a next run cycle for a target machine device based on target sound characteristics may include:
Step S131: and performing fault prediction of the next operation period on the target machine equipment by adopting the sound characteristics sensitive to the faults of the target machine equipment in the plurality of sound characteristics.
The embodiment of step S131 described above is, for example: it will be appreciated that, since the sound features that are sensitive to the failure of the target machine device have been determined during the training process of the machine learning model, the sound features that are sensitive to the failure of the target machine device may be directly screened from the plurality of sound features, i.e., the screened sound features may be obtained. And then, according to the screened sound characteristics, carrying out fault prediction of the next operation period on the target machine equipment.
The fault-sensitive sound characteristic may be determined based on characteristics (e.g., material, sound characteristics, etc.) of the sound-emitting component of the target machine device, such as the motor being different from the fault-sensitive sound characteristic of the track. Fault-sensitive acoustic features may also be determined by analyzing the sample data during the machine learning model training phase.
It should be understood that, in combination with the foregoing solutions of step S123-step S124, the filtering of the fault-sensitive sound features herein may filter the corresponding fault-sensitive sound features for different institutions of the target machine device, and the fault-sensitive sound features filtered by the different institutions may be different.
As an alternative embodiment of the above step S130, an embodiment of performing the fault prediction of the next operation cycle on the target machine device according to the target sound feature may include:
Step S132: a health index of the target machine device is determined based on the target sound characteristics.
The embodiment of step S132 described above is, for example: it will be appreciated that the target sound characteristic may be a fault-sensitive multiple sound characteristic of the target machine device, and the formula may be usedPerforming health assessment on a plurality of sound features of fault sensitivity of target machine equipment to obtain a health index; wherein/>Is an obtained health index, which can be set in the range of 0-100,/>Representing each of the filtered feature weight parameters, which may be summarized by human analysis (e.g., selected based on expert experience),/>, for exampleRepresenting the real-time value of the characteristic parameter.
Step S133: and carrying out fault prediction of the next operation period on the target machine equipment according to the health index.
The embodiment of step S133 described above is, for example: judging whether the health index is smaller than a preset threshold value, and if the health index is smaller than the preset threshold value, determining that the target machine equipment has faults; similarly, if the health index is greater than or equal to a preset threshold, determining that the target machine equipment has no fault. There are a number of situations where a fault exists, which may be determined based on the specific configuration of the target machine device, and this process is described in detail below.
Alternatively, the staff may view the health index of the target machine devices accessed to the system in an information management system (e.g., health management platform system), although the staff may also view the current operating state and health index of each target machine device remotely through the information management system (e.g., health management platform system). If the health index is smaller than a preset threshold (for example, the preset alarm value is 60), or if the health index has a long-term trend, the generation of the maintenance work order can be automatically triggered, and the maintenance work order is sent to the corresponding maintenance personnel so as to remind the maintenance personnel of timely maintenance and repair. If the health index is greater than or equal to the preset threshold (for example, the preset alarm value is 60), the time for carrying out the abnormality prediction next time can be set according to the health index.
In the implementation process of the scheme, the fault prediction of the next operation period is carried out on the target machine equipment according to the health index of the target machine equipment determined by the sound characteristics, so that the fault of the target machine equipment is effectively quantized into the health index, and the accuracy of the fault prediction is improved. In some scenes, the health management can be performed on the equipment according to the health index, so that the maintenance and management capacity of the equipment is improved.
As an alternative embodiment of the above step S130, after determining the health index of the target machine device according to the target sound characteristic, it may further include:
step S134: and if the health index is in the alarm value interval, generating alarm information.
The alarm value interval refers to an interval which is preset and used for alarm, and the interval can be set according to specific situations, for example: the preset alarm value interval is set to 60-80.
The embodiment of step S134 described above is, for example: if the health index is 70, the health index is located in the alarm value interval (for example, the preset alarm value interval is 60-80), the prediction result is determined to be abnormal, and then the generation of an alarm work order (the alarm work order is one of alarm information) can be automatically triggered and sent to corresponding maintenance personnel to remind the maintenance personnel of timely maintenance and repair.
It should be appreciated that the boundaries of the alert value interval may be different from the health index value identifying the equipment failure, e.g., although the health index of the machine equipment is a higher value than the machine equipment failure alert, because of the closer proximity to the machine equipment failure alert, indicating that the target machine equipment may be operating poorly, alert information may still be generated to alert personnel.
Optionally, the alarm information may further include a device health degree information, for example: severe faults, light faults, poor operation, etc.
Please refer to fig. 3, which is a schematic flow chart of a stacker failure prediction method according to an embodiment of the present application; the stacker fault prediction method can be used for predicting faults of the next operation period of the stacker, the stacker comprises a horizontal operation mechanism, a lifting operation mechanism and a fork telescopic mechanism, the fork telescopic mechanism is arranged on the lifting operation mechanism, namely, the fork telescopic mechanism can be arranged aiming at the lifting operation mechanism, and the lifting operation mechanism can be arranged aiming at the horizontal operation mechanism. Embodiments of the above method may include:
step S210: and acquiring sound data to be processed, wherein the sound data to be processed is sound data of a horizontal running mechanism, a lifting running mechanism and a fork telescopic mechanism in the acquired stacker in one running period.
For example: in one operation period, the horizontal motor, the gear rack, the top rail, the bottom rail and the like of the horizontal operation mechanism can possibly generate sound. In one operation period, the lifting motor, the chain wheel and the chain and the like of the lifting operation mechanism can possibly generate sound. In one operation period, the telescopic motor, the telescopic plate and the like of the fork telescopic mechanism can possibly generate sound. These sounds may be collected by a sound sensor.
Step S220: and predicting target sound characteristics of the stacker in the next operation period according to the sound data to be processed.
Step S230: and carrying out fault prediction on the mechanism of the stacker in the next operation period according to the target sound characteristics.
The implementation principle and implementation of the steps S210 to S230 are similar to those of the steps S110 to S130, and thus, the implementation principle and implementation thereof will not be described herein, and reference may be made to the description of the steps S110 to S130.
As an alternative embodiment of the above step S220, an embodiment of predicting the target sound characteristic of the stacker in the next operation cycle according to the sound data to be processed may include:
step S221: and extracting sound characteristics of the sound data to be processed to obtain the sound characteristics to be processed.
Step S222: predicting the sound characteristics to be processed through a machine learning model to obtain target sound characteristics; wherein the machine learning model is trained from sample data comprising: the sound characteristics of sound data of the stacker in one operation cycle and the sound characteristics of sound data of the stacker in the next operation cycle of the operation cycle.
The implementation principle and implementation of this step S221 to step S222 are similar to those of the step S121 to step S122, and thus, the implementation principle and implementation thereof will not be described again, and reference may be made to the description of the step S121 to step S122.
As an alternative embodiment of the step S222, the type of the target sound feature may include at least one of a time domain feature, a frequency domain feature, and a time frequency feature. When the target sound characteristics are predicted through the machine learning model, one type of sound characteristics to be processed is input, and the same type of target sound characteristics can be obtained. For example: if the type of the target sound feature comprises the time domain feature, predicting the to-be-processed sound feature of the time domain feature type through a machine learning model to obtain the target sound feature of the time domain feature type. If the type of the target sound feature comprises the frequency domain feature, predicting the sound feature to be processed of the frequency domain feature type through a machine learning model to obtain the target sound feature of the frequency domain feature type. If the type of the target sound feature comprises the time-frequency feature, predicting the to-be-processed sound feature of the time-frequency feature type through a machine learning model to obtain the target sound feature of the time-frequency feature type.
As an alternative embodiment of the above step S220, an embodiment of predicting the target sound characteristic of the stacker in the next operation cycle according to the sound data to be processed may include:
step S223: dividing the sound data to be processed into the sound data to be processed with different working conditions according to the control signal time sequence data of the stacker.
It can be understood that, because the time and/or frequency of the sound emitted by the stacker under different working conditions such as the horizontal sliding rail working condition, the vertical sliding rail working condition, the fork telescopic working condition and the like are greatly different, the sound data to be processed can be classified according to the control signal time sequence data of the stacker, i.e. the sound data to be processed can be divided into the sound data to be processed under different working conditions according to the control signal time sequence data of the stacker.
Step S224: and predicting target sound characteristics in the next operation period of each working condition according to the sound data to be processed of the working condition.
Wherein, different operating modes include: the horizontal running mechanism running condition, the lifting running mechanism running condition and the fork telescopic mechanism running condition; aiming at the target sound characteristics predicted by the sound data to be processed of a working condition, carrying out fault prediction of the next operation period on a mechanism of the stacker in an operation state in the working condition.
The implementation principle and implementation of this step S223 to step S224 are similar to those of the step S123 to step S124, and thus, the implementation principle and implementation thereof will not be described again here, and reference may be made to the description of the step S123 to step S124.
As an alternative embodiment of the above step S223, the embodiment of dividing the to-be-processed sound data into to-be-processed sound data of different working conditions according to the control signal time sequence data of the stacker may include:
step S223a: searching acquisition time meeting a first preset condition from control signal time sequence data of the stacker, wherein the first preset condition comprises the following steps: the fork telescoping mechanism is in a motion state during the run cycle, and the horizontal run mechanism and the lift run mechanism are in a stationary state during the run cycle.
Step S223b: and searching sound data meeting the first preset condition at the acquisition time from the sound data to be processed.
Step S223c: and dividing the sound data meeting the first preset condition at the acquisition time into sound data of the operation working condition of the fork telescopic mechanism.
The embodiments of the above steps S223a to S223c are, for example: referring to fig. 2, in 10 th to 15 th seconds of the control signal time sequence data, the positions of the stacker on the X axis, the Z axis and the fork axis do not move, which means that the fork telescopic mechanism is in a moving state in the operating cycle, and the horizontal operating mechanism and the lifting operating mechanism are in a stationary state in the operating cycle, so that the control signal time sequence data of the 10 th to 15 th seconds of the part of the stacker is the acquisition time satisfying the first preset condition, and then the sound data of the acquisition time satisfying the first preset condition can be divided into the sound data of the operating condition of the fork telescopic mechanism.
As an alternative embodiment of the above step S223, the embodiment of dividing the to-be-processed sound data into to-be-processed sound data of different working conditions according to the control signal time sequence data of the stacker may include:
Step S223d: searching acquisition time meeting a second preset condition from control signal time sequence data of the stacker, wherein the second preset condition comprises the following steps: the fork telescoping mechanism is in a stationary state during the run cycle, and the horizontal run mechanism and the lift run mechanism are in a moving state during the run cycle.
Step S223e: and searching sound data meeting the second preset condition at the acquisition time from the sound data to be processed.
Step S223f: according to the frequency domain characteristic difference of the sound data when each mechanism of the stacker operates independently, the sound data meeting the second preset condition at the acquisition time is divided into the sound data of the operating condition of the horizontal operating mechanism and the sound data of the operating condition of the lifting operating mechanism.
The embodiments of the steps S223d to S223f are similar to those of the steps S223a to S223c, and thus are not described in detail.
As an alternative embodiment of the step S230, the target sound feature may include a plurality of sound features, and the embodiment of predicting the failure of the mechanism of the stacker according to the target sound feature may include:
Step S231: and respectively carrying out fault prediction on each mechanism of the stacker by adopting the sound characteristics which are sensitive to faults of each mechanism of the stacker in the plurality of sound characteristics.
The embodiment of step S231 is similar to that of step S131, and thus will not be described in detail.
As an alternative implementation manner of the stacker fault prediction method, the sound characteristics sensitive to faults of each mechanism of the stacker may be obtained by screening by using a principal component analysis method. The embodiment is similar to the embodiment of the optional embodiment of the equipment failure prediction method using the principal component analysis method for screening, and therefore will not be described in detail.
As an alternative embodiment of the above step S230, performing fault prediction on the mechanism of the stacker for the next operation cycle according to the target sound feature includes:
Step S232: and determining health indexes of various institutions of the stacker according to the target sound characteristics.
Step S233: and carrying out fault prediction on each mechanism of the stacker for the next operation period according to the health index.
The implementation principle and implementation of this step S232 to step S233 are similar to those of the step S132 to step S133, and thus, the implementation principle and implementation thereof will not be described again here, and reference may be made to the description of the step S132 to step S133.
As an alternative embodiment of the above stacker fault prediction method, after determining the health index of each organization of the stacker according to the target sound feature, the method may further include:
step S234: and if the health index is in the alarm value interval, generating alarm information.
The embodiment of step S234 is similar to that of step S134, and thus will not be described again.
The embodiment of the application also provides a device fault prediction system, which comprises: a sound sensor and a computing device.
The sound sensor is used for collecting sound data to be processed, wherein the sound data to be processed is the sound data of the target machine equipment in one operation period.
And the computing device is used for predicting the target sound characteristics of the target machine equipment in the next operation period according to the sound data to be processed and carrying out fault prediction of the target machine equipment in the next operation period according to the target sound characteristics.
Optionally, in an embodiment of the present application, the computing device includes an edge device and a server.
The edge device is used for: and extracting sound characteristics of the sound data to be processed to obtain the sound characteristics to be processed.
The server is used for: predicting the sound characteristics to be processed through a machine learning model to obtain target sound characteristics.
And carrying out fault prediction of the next operation period on the target machine equipment according to the target sound characteristics.
Wherein, the machine learning model is formed according to sample data training, and the sample data includes: the sound characteristics of the sound data of the target machine device in one operation cycle and the sound characteristics of the sound data of the target machine device in the next operation cycle of the operation cycle.
Optionally, in the embodiment of the present application, the edge device is specifically configured to:
And dividing the sound data to be processed into the sound data to be processed with different working conditions according to the control signal time sequence data of the target machine equipment.
And extracting sound characteristics of the sound data to be processed in each working condition to obtain the sound characteristics to be processed.
The method comprises the steps of predicting a target sound characteristic predicted by sound data to be processed under a working condition, and predicting a fault of a next operation period of a mechanism of target machine equipment in an operation state under the working condition.
Optionally, in the embodiment of the present application, the edge device is specifically configured to:
dividing sound data of a first preset working condition of target machine equipment from the sound data to be processed according to the control signal time sequence data;
dividing the sound data of the second preset working condition and the sound data of the third preset working condition of the target machine equipment from the sound data except the sound data of the first preset working condition in the sound data to be processed according to the frequency domain characteristic difference of the sound data when all mechanisms of the target machine equipment independently operate;
The target machine equipment is different in mechanisms in running states in a first preset working condition, a second preset working condition and a third preset working condition.
Optionally, in an embodiment of the present application, the target sound feature includes a plurality of sound features.
The server is specifically configured to: and performing fault prediction of the next operation period on the target machine equipment by adopting the sound characteristics sensitive to the faults of the target machine equipment in the plurality of sound characteristics.
Optionally, in an embodiment of the present application, the server is specifically configured to:
a health index of the target machine device is determined based on the target sound characteristics.
And carrying out fault prediction of the next operation period on the target machine equipment according to the health index.
Optionally, in an embodiment of the present application, the server is further configured to:
and if the health index is in the alarm value interval, generating alarm information.
The embodiments of the components in the above-mentioned equipment failure prediction system have been described in the foregoing equipment failure prediction method or stacker failure prediction method, and will not be repeated.
Please refer to fig. 4, which illustrates a schematic structural diagram of a stacker provided in an embodiment of the present application; embodiments of the present application provide a stacker 300 that may include: a horizontal running mechanism 310, a lifting running mechanism 320, and a fork telescoping mechanism 330, and an acoustic sensor;
The horizontal running mechanism 310 may include: the upper top rail 311, the lower ground rail 313 and the horizontal rack 312 are movably connected with the gear of the top rail 311 or the gear of the ground rail 313, so that the rack 312 can be driven to slide horizontally on the top rail 311 and the ground rail 313 of the stacker 300.
The lifting mechanism 320 is disposed on the horizontal running mechanism, that is, the lifting mechanism is disposed for the horizontal running mechanism, and the lifting mechanism may be movably connected with the horizontal running mechanism so as to perform lifting movement and horizontal movement at the same time. The elevation operation mechanism 320 includes: the lifting motor 321 and the sprocket chain 322 are movably connected with the gear or the sprocket of the fork telescopic mechanism 330, so that the lifting motor 321 can drive the sprocket chain 322, thereby enabling the fork telescopic mechanism 330 of the stacker 300 to vertically lift.
The fork telescopic mechanism is arranged on the lifting operation mechanism; the fork telescoping mechanism 330 includes: the telescopic plate 331 and the telescopic motor 332, the telescopic motor 332 and the telescopic plate 331 can be movably connected, and therefore, the telescopic motor 332 can drive the telescopic plate 331 to perform telescopic motion.
In replacing batteries for electric vehicles using a stacker, the battery to be replaced may be removed from the battery warehouse using the stacker and then placed on an automated guided vehicle (Automated Guided Vehicle, AGV) or a guided vehicle (Rail Guided Vehicle, RGV). The specific process is as follows: the telescopic motor in the fork telescopic machanism can drive the expansion plate and carry out telescopic motion for the expansion plate can take out the battery from the goods shelves of battery warehouse, then, drive sprocket chain through elevator motor, thereby make the battery on the expansion plate carry out vertical lifting motion, then, drive rack and pinion of horizontal direction by the motor, make the stacker carry out horizontal motion (usually rectilinear motion) on top of the earth rail and rail, finally, transport and place the battery on Automated Guided Vehicle (AGV) or have Rail Guided Vehicle (RGV) by the stacker.
The sound sensor is arranged at least one of the horizontal running mechanism, the lifting running mechanism and the fork telescopic mechanism and is used for collecting sound data when the stacker runs. The number of the sound sensors may be only one, for example, one sound sensor may be provided at a position of one of the three mechanisms or at a relatively intermediate position of the three mechanisms. There may be more sound sensors, different sound sensors being arranged at different locations, e.g. one or more sound sensors for each of the above-mentioned mechanisms, respectively.
Alternatively, when a sound sensor is provided for each mechanism of the stacker, the failure prediction of the mechanism may be performed using the sound data to be processed collected by the sensor provided for each mechanism.
Optionally, in an embodiment of the present application, the sound sensor includes a first part of the sound sensor disposed on the horizontal running mechanism; the first partial sound sensor is arranged at the position of a part to be monitored of the adjacent horizontal running mechanism, and the part to be monitored comprises: at least one of a horizontal motor, a gear rack, a top rail and a bottom rail.
Optionally, in an embodiment of the present application, the sound sensor includes a second part of the sound sensor disposed on the lifting operation mechanism; the second partial sound sensor is arranged at the position of a part to be monitored of the adjacent lifting operation mechanism, and the part to be monitored comprises: at least one of a lift motor and a sprocket chain.
Optionally, in an embodiment of the present application, the sound sensor includes a third part of the sound sensor disposed on the fork telescopic mechanism; the third part sound sensor sets up in the position of the part of waiting to monitor of close fork telescopic machanism, waits to monitor the part and includes: at least one of a telescoping motor and a telescoping plate.
Please refer to fig. 5, which illustrates a schematic structure of a computing device according to an embodiment of the present application. A computing device 400 provided by an embodiment of the present application includes: a processor 410 and a memory 420, the memory 420 storing machine-readable instructions executable by the processor 410, which when executed by the processor 410 perform the method as described above.
The embodiment of the present application also provides a computer readable storage medium 430, on which computer readable storage medium 430 a computer program is stored which, when executed by the processor 410, performs a method as above. The computer-readable storage medium 430 may be implemented by any type or combination of volatile or non-volatile Memory devices, such as static random access Memory (Static Random Access Memory, SRAM for short), electrically erasable programmable Read-Only Memory (ELECTRICALLY ERASABLE PROGRAMMABLE READ-Only Memory, EEPROM for short), erasable programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM for short), programmable Read-Only Memory (Programmable Read-Only Memory, PROM for short), read-Only Memory (ROM for short), magnetic Memory, flash Memory, magnetic disk, or optical disk.
It should be noted that, in the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described as different from other embodiments, and identical and similar parts between the embodiments are all enough to be referred to each other. For the apparatus class embodiments, the description is relatively simple as it is substantially similar to the method embodiments, and reference is made to the description of the method embodiments for relevant points.
In the embodiments of the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The apparatus embodiments described above are merely illustrative, for example, of the flowcharts and block diagrams in the figures that illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.
In addition, the functional modules of the embodiments of the present application may be integrated together to form a single part, or the modules may exist separately, or two or more modules may be integrated to form a single part. Furthermore, in the description herein, the descriptions of the terms "one embodiment," "some embodiments," "examples," "specific examples," "some examples," and the like are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the embodiments of the present application. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.
The foregoing description is merely an optional implementation of the embodiment of the present application, but the scope of the embodiment of the present application is not limited thereto, and any person skilled in the art may easily think about changes or substitutions within the technical scope of the embodiment of the present application, and the changes or substitutions are covered by the scope of the embodiment of the present application.
Claims (20)
1. The fault prediction method for the stacker is characterized in that the stacker comprises a horizontal running mechanism, a lifting running mechanism and a fork telescopic mechanism, wherein the fork telescopic mechanism is arranged on the lifting running mechanism, and the lifting running mechanism is arranged on the horizontal running mechanism; the method comprises the following steps:
Acquiring sound data to be processed, wherein the sound data to be processed is the sound data of the horizontal running mechanism, the lifting running mechanism and the fork telescopic mechanism in the acquired stacker in one running period; wherein, the one operation period is a process of taking goods and storing inventory for the stacker;
Predicting target sound characteristics of the stacker in the next operation period according to the sound data to be processed;
Performing fault prediction on a mechanism of the stacker according to the target sound characteristics;
The predicting the target sound characteristic of the stacker in the next operation period according to the sound data to be processed comprises the following steps: predicting the sound data to be processed or the sound characteristics extracted according to the sound data to be processed through a machine learning model to obtain target sound characteristics of the stacker in a next operation period, wherein the machine learning model is formed according to sample data training, and the sample data comprises: the sound data with the change trend of the stacker in one operation period and in the next operation period of the operation period, or the sound characteristics obtained by extracting the sound data with the change trend of the stacker in one operation period and in the next operation period of the operation period;
The predicting the target sound characteristic of the stacker in the next operation period according to the sound data to be processed comprises the following steps: dividing the sound data to be processed into the sound data to be processed of different working conditions of the stacker according to the control signal time sequence data of the stacker; dividing sound data of a first preset working condition of the stacker from the sound data to be processed according to the control signal time sequence data of the stacker; dividing sound data of a second preset working condition and sound data of a third preset working condition of the stacker from sound data except the sound data of the first preset working condition in the sound data to be processed according to frequency domain characteristic differences of the sound data when all mechanisms of the stacker independently operate; wherein, different operating modes include: the horizontal running mechanism running condition, the lifting running mechanism running condition and the fork telescopic mechanism running condition; and aiming at the target sound characteristics predicted by the sound data to be processed of a working condition, carrying out fault prediction on a mechanism of the stacker in an operating state in the working condition.
2. The method of claim 1, wherein predicting a target sound characteristic of the stacker in a next operation cycle based on the sound data to be processed comprises:
Extracting sound characteristics of the sound data to be processed to obtain sound characteristics to be processed;
and predicting the sound characteristics to be processed through a machine learning model to obtain the target sound characteristics.
3. The method of claim 2, wherein the type of target sound feature comprises at least one of a time domain feature, a frequency domain feature, and a time frequency feature;
If the type of the target sound feature includes a time domain feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including:
predicting the to-be-processed sound features of the time domain feature type through a machine learning model to obtain the target sound features of the time domain feature type;
if the type of the target sound feature includes a frequency domain feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including:
Predicting the to-be-processed sound features of the frequency domain feature type through a machine learning model to obtain the target sound features of the frequency domain feature type;
if the type of the target sound feature includes a time-frequency feature, predicting the sound feature to be processed through a machine learning model to obtain the target sound feature, including:
predicting the to-be-processed sound features of the time-frequency feature type through a machine learning model to obtain the target sound features of the time-frequency feature type.
4. The method according to claim 1, wherein the dividing the sound data to be processed into the sound data to be processed of different working conditions of the stacker according to the control signal timing data of the stacker comprises:
Searching acquisition time meeting a first preset condition from control signal time sequence data of the stacker, wherein the first preset condition comprises the following steps: the fork telescopic mechanism is in a motion state in the operation period, and the horizontal operation mechanism and the lifting operation mechanism are in a static state in the operation period;
Searching sound data meeting the first preset condition at the acquisition time from the sound data to be processed;
And dividing the sound data meeting the first preset condition at the acquisition time into the sound data of the operation working condition of the fork telescopic mechanism.
5. The method according to claim 1, wherein the dividing the sound data to be processed into the sound data to be processed of different working conditions of the stacker according to the control signal timing data of the stacker comprises:
Searching acquisition time meeting a second preset condition from control signal time sequence data of the stacker, wherein the second preset condition comprises the following steps: the fork telescopic mechanism is in a static state in the running period, and the horizontal running mechanism and the lifting running mechanism are in a moving state in the running period;
searching sound data meeting the second preset condition at the acquisition time from the sound data to be processed;
According to the frequency domain characteristic difference of the sound data when the horizontal running mechanism and the lifting running mechanism independently run, the sound data meeting the second preset condition at the acquisition time is divided into the sound data of the running working condition of the horizontal running mechanism and the sound data of the running working condition of the lifting running mechanism.
6. The method of any one of claims 1 to 5, wherein the target sound feature comprises a plurality of sound features;
The fault prediction for the mechanism of the stacker according to the target sound characteristics comprises the following steps:
And respectively predicting faults of all the mechanisms of the stacker by adopting the sound characteristics which are sensitive to faults of all the mechanisms of the stacker in the sound characteristics.
7. The method of claim 1, wherein said predicting a malfunction of a mechanism of the stacker based on the target sound signature comprises:
Determining health indexes of all institutions of the stacker according to the target sound characteristics;
And carrying out fault prediction on each mechanism of the stacker according to the health index.
8. The method of claim 7, further comprising, after said determining health index for each institution of the stacker based on the target sound signature:
and if the health index is positioned in the alarm numerical value interval, generating alarm information.
9. A stacker fault prediction system, comprising: a sound sensor and a computing device;
The sound sensor is used for collecting sound data to be processed, wherein the sound data to be processed is sound data of target machine equipment in one operation period; the target machine equipment is a stacker, and the stacker comprises a horizontal running mechanism, a lifting running mechanism and a fork telescopic mechanism, wherein the fork telescopic mechanism is arranged on the lifting running mechanism, and the lifting running mechanism is arranged on the horizontal running mechanism; the sound data to be processed are collected sound data of the horizontal running mechanism, the lifting running mechanism and the fork telescopic mechanism in the stacker in a running period; wherein, the one operation period is a process of taking goods and storing inventory for the stacker;
The computing device is used for predicting target sound characteristics of the target machine device in the next operation period according to the sound data to be processed;
Performing fault prediction on the target machine equipment according to the target sound characteristics;
Wherein the predicting the target sound characteristic of the target machine device in the next operation period according to the sound data to be processed includes: predicting the sound data to be processed or the sound characteristics extracted according to the sound data to be processed through a machine learning model to obtain target sound characteristics of the target machine equipment in a next operation period, wherein the machine learning model is formed according to sample data training, and the sample data comprises: the target machine equipment extracts sound characteristics obtained according to the sound data with the change trend of the target machine equipment in one operation period and the next operation period of the operation period;
The computing device includes an edge device, the edge device specifically configured to: dividing sound data of a first preset working condition of the stacker from the sound data to be processed according to the control signal time sequence data of the stacker; dividing sound data of a second preset working condition and sound data of a third preset working condition of the stacker from sound data except the sound data of the first preset working condition in the sound data to be processed according to frequency domain characteristic differences of the sound data when all mechanisms of the stacker independently operate; wherein, different operating modes include: the horizontal running mechanism running condition, the lifting running mechanism running condition and the fork telescopic mechanism running condition; and aiming at the target sound characteristics predicted by the sound data to be processed of a working condition, carrying out fault prediction on a mechanism of the stacker in an operating state in the working condition.
10. The system of claim 9, wherein the computing device further comprises a server;
The edge device is used for: extracting sound characteristics of the sound data to be processed to obtain sound characteristics to be processed;
the server is used for: predicting the sound characteristics to be processed through a machine learning model to obtain the target sound characteristics;
Performing fault prediction on the target machine equipment according to the target sound characteristics;
Wherein the machine learning model is trained from sample data comprising: the sound characteristics of the sound data of the target machine device in one operation period and the sound characteristics of the sound data of the target machine device in the next operation period of the operation period.
11. The system according to claim 9, wherein the edge device is specifically configured to:
Searching acquisition time meeting a first preset condition from control signal time sequence data of the stacker, wherein the first preset condition comprises the following steps: the fork telescopic mechanism is in a motion state in the operation period, and the horizontal operation mechanism and the lifting operation mechanism are in a static state in the operation period; searching sound data meeting the first preset condition at the acquisition time from the sound data to be processed; dividing sound data at the acquisition time meeting the first preset condition into sound data of the operation working condition of the fork telescopic mechanism; and/or
Searching acquisition time meeting a second preset condition from control signal time sequence data of the stacker, wherein the second preset condition comprises the following steps: the fork telescopic mechanism is in a static state in the running period, and the horizontal running mechanism and the lifting running mechanism are in a moving state in the running period; searching sound data meeting the second preset condition at the acquisition time from the sound data to be processed; according to the frequency domain characteristic difference of the sound data when the horizontal running mechanism and the lifting running mechanism independently run, the sound data meeting the second preset condition at the acquisition time is divided into the sound data of the running working condition of the horizontal running mechanism and the sound data of the running working condition of the lifting running mechanism.
12. The system of claim 10, wherein the target sound feature comprises a plurality of sound features;
the server is specifically configured to: and predicting the fault of the target machine equipment by adopting the sound characteristics which are sensitive to the fault of the target machine equipment in the target sound characteristics.
13. The system according to claim 10, characterized in that the server is in particular adapted to:
determining a health index of the target machine device according to the target sound characteristics;
and carrying out fault prediction on the target machine equipment according to the health index.
14. The system of claim 13, wherein the server is further configured to:
and if the health index is positioned in the alarm numerical value interval, generating alarm information.
15. A stacker, comprising: the device comprises a horizontal running mechanism, a lifting running mechanism, a fork telescopic mechanism and a sound sensor, wherein the fork telescopic mechanism is arranged on the lifting running mechanism;
The sound sensor is arranged at least one of the horizontal running mechanism, the lifting running mechanism and the fork telescopic mechanism, and is used for collecting sound data to be processed, wherein the sound data to be processed is the sound data of the stacker in one running period, and the sound data to be processed is used for carrying out fault prediction on the stacker through the method of any one of claims 1 to 8.
16. The stacker of claim 15 wherein said acoustic sensor comprises a first portion of an acoustic sensor disposed on said horizontal running mechanism; the first partial sound sensor is arranged at a position close to a part to be monitored of the horizontal running mechanism, and the part to be monitored comprises: at least one of a horizontal motor, a gear rack, a top rail and a bottom rail.
17. The stacker of claim 15 wherein said acoustic sensor comprises a second portion acoustic sensor disposed on said lift mechanism; the second part sound sensor is arranged at a position close to a part to be monitored of the lifting operation mechanism, and the part to be monitored comprises: at least one of a lift motor and a sprocket chain.
18. The stacker of claim 15 wherein said acoustic sensor comprises a third portion acoustic sensor disposed on said fork telescoping mechanism; the third part sound sensor is arranged at a position close to a part to be monitored of the fork telescopic mechanism, and the part to be monitored comprises: at least one of a telescoping motor and a telescoping plate.
19. A computing device comprising a processor and a memory storing computer readable instructions that, when executed by the processor, perform the method of any of claims 1-8.
20. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, performs the method according to any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310673692.5A CN116403605B (en) | 2023-06-08 | Stacker fault prediction method and related device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310673692.5A CN116403605B (en) | 2023-06-08 | Stacker fault prediction method and related device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116403605A CN116403605A (en) | 2023-07-07 |
| CN116403605B true CN116403605B (en) | 2024-06-07 |
Family
ID=
Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN205873020U (en) * | 2016-07-19 | 2017-01-11 | 国网天津市电力公司 | Stacker centre gripping sorting device |
| CN110107461A (en) * | 2019-05-22 | 2019-08-09 | 华润新能源(太原)有限公司 | Fan trouble method for early warning, device, equipment and computer readable storage medium |
| CN110346157A (en) * | 2018-04-04 | 2019-10-18 | 国网安徽省电力有限公司电力科学研究院 | A kind of application method that wavelet singular entropy is detected in piler cyclic breakdown |
| CN110940539A (en) * | 2019-12-03 | 2020-03-31 | 桂林理工大学 | Machine equipment fault diagnosis method based on artificial experience and voice recognition |
| CN111170214A (en) * | 2020-01-20 | 2020-05-19 | 广东垒亚安防科技有限公司 | Double-fork double-extension fork stacker |
| WO2020153934A1 (en) * | 2019-01-21 | 2020-07-30 | Hewlett-Packard Development Company, L.P. | Fault prediction model training with audio data |
| CN112633492A (en) * | 2019-10-08 | 2021-04-09 | 佛山市顺德区美的电热电器制造有限公司 | Failure processing method, failure processing device, information processing device, storage medium, and server |
| CN213445138U (en) * | 2020-10-28 | 2021-06-15 | 深圳市纳瑞自动化设备有限公司 | Tunnel stacker |
| CN113140229A (en) * | 2021-04-21 | 2021-07-20 | 上海泛德声学工程有限公司 | Sound detection method based on neural network, industrial acoustic detection system and method |
| CN113205829A (en) * | 2021-04-29 | 2021-08-03 | 上海智大电子有限公司 | Method and system for comprehensively monitoring running state of equipment |
| CN113409818A (en) * | 2021-08-18 | 2021-09-17 | 宁波帅特龙集团有限公司 | Predictive maintenance method and system for production equipment based on sound sensor |
| CN113470694A (en) * | 2021-04-25 | 2021-10-01 | 重庆市科源能源技术发展有限公司 | Remote listening monitoring method, device and system for hydraulic turbine set |
| CN113493083A (en) * | 2021-07-30 | 2021-10-12 | 因格(苏州)智能技术有限公司 | Light stereoscopic warehouse and automatic material storing and taking method |
| CN114372491A (en) * | 2021-12-31 | 2022-04-19 | 山东科讯信息科技有限公司 | Fault identification method of rotating equipment and related equipment |
| CN114528942A (en) * | 2022-02-23 | 2022-05-24 | 上海华兴数字科技有限公司 | Construction method of data sample library of engineering machinery, failure prediction method and engineering machinery |
| CN114722866A (en) * | 2022-03-28 | 2022-07-08 | 杭州中网智慧科技有限公司 | Management and control system is tracked to equipment based on industry internet |
| CN115165362A (en) * | 2022-06-24 | 2022-10-11 | 鞍钢集团矿业有限公司 | Intelligent bearing fault detection system and method based on inspection and vibration signals |
| CN115169505A (en) * | 2022-09-06 | 2022-10-11 | 杭州浅水数字技术有限公司 | Early warning method and early warning system for mechanical fault of special equipment moving part |
| CN115186904A (en) * | 2022-07-13 | 2022-10-14 | 清华大学 | Industrial equipment fault prediction method and device based on Transformer |
| CN115406630A (en) * | 2021-05-27 | 2022-11-29 | 循声科技(重庆)有限公司 | Method for detecting faults of wind driven generator blades through passive acoustic signals based on machine learning |
| CN115409094A (en) * | 2022-08-17 | 2022-11-29 | 中建材智慧工业科技有限公司 | Equipment fault prediction method, device and storage medium |
| CN115452392A (en) * | 2021-11-30 | 2022-12-09 | 广州汽车集团股份有限公司 | Acoustic diagnosis method and system for vehicle engine fault |
Patent Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN205873020U (en) * | 2016-07-19 | 2017-01-11 | 国网天津市电力公司 | Stacker centre gripping sorting device |
| CN110346157A (en) * | 2018-04-04 | 2019-10-18 | 国网安徽省电力有限公司电力科学研究院 | A kind of application method that wavelet singular entropy is detected in piler cyclic breakdown |
| WO2020153934A1 (en) * | 2019-01-21 | 2020-07-30 | Hewlett-Packard Development Company, L.P. | Fault prediction model training with audio data |
| CN110107461A (en) * | 2019-05-22 | 2019-08-09 | 华润新能源(太原)有限公司 | Fan trouble method for early warning, device, equipment and computer readable storage medium |
| CN112633492A (en) * | 2019-10-08 | 2021-04-09 | 佛山市顺德区美的电热电器制造有限公司 | Failure processing method, failure processing device, information processing device, storage medium, and server |
| CN110940539A (en) * | 2019-12-03 | 2020-03-31 | 桂林理工大学 | Machine equipment fault diagnosis method based on artificial experience and voice recognition |
| CN111170214A (en) * | 2020-01-20 | 2020-05-19 | 广东垒亚安防科技有限公司 | Double-fork double-extension fork stacker |
| CN213445138U (en) * | 2020-10-28 | 2021-06-15 | 深圳市纳瑞自动化设备有限公司 | Tunnel stacker |
| CN113140229A (en) * | 2021-04-21 | 2021-07-20 | 上海泛德声学工程有限公司 | Sound detection method based on neural network, industrial acoustic detection system and method |
| CN113470694A (en) * | 2021-04-25 | 2021-10-01 | 重庆市科源能源技术发展有限公司 | Remote listening monitoring method, device and system for hydraulic turbine set |
| CN113205829A (en) * | 2021-04-29 | 2021-08-03 | 上海智大电子有限公司 | Method and system for comprehensively monitoring running state of equipment |
| CN115406630A (en) * | 2021-05-27 | 2022-11-29 | 循声科技(重庆)有限公司 | Method for detecting faults of wind driven generator blades through passive acoustic signals based on machine learning |
| CN113493083A (en) * | 2021-07-30 | 2021-10-12 | 因格(苏州)智能技术有限公司 | Light stereoscopic warehouse and automatic material storing and taking method |
| CN113409818A (en) * | 2021-08-18 | 2021-09-17 | 宁波帅特龙集团有限公司 | Predictive maintenance method and system for production equipment based on sound sensor |
| CN115452392A (en) * | 2021-11-30 | 2022-12-09 | 广州汽车集团股份有限公司 | Acoustic diagnosis method and system for vehicle engine fault |
| CN114372491A (en) * | 2021-12-31 | 2022-04-19 | 山东科讯信息科技有限公司 | Fault identification method of rotating equipment and related equipment |
| CN114528942A (en) * | 2022-02-23 | 2022-05-24 | 上海华兴数字科技有限公司 | Construction method of data sample library of engineering machinery, failure prediction method and engineering machinery |
| CN114722866A (en) * | 2022-03-28 | 2022-07-08 | 杭州中网智慧科技有限公司 | Management and control system is tracked to equipment based on industry internet |
| CN115165362A (en) * | 2022-06-24 | 2022-10-11 | 鞍钢集团矿业有限公司 | Intelligent bearing fault detection system and method based on inspection and vibration signals |
| CN115186904A (en) * | 2022-07-13 | 2022-10-14 | 清华大学 | Industrial equipment fault prediction method and device based on Transformer |
| CN115409094A (en) * | 2022-08-17 | 2022-11-29 | 中建材智慧工业科技有限公司 | Equipment fault prediction method, device and storage medium |
| CN115169505A (en) * | 2022-09-06 | 2022-10-11 | 杭州浅水数字技术有限公司 | Early warning method and early warning system for mechanical fault of special equipment moving part |
Non-Patent Citations (2)
| Title |
|---|
| 李杰.工业人工智能.上海交通大学出版社,2019,129-141. * |
| 谭洪卫.图说公共建筑能耗的数据挖掘与模型方法.同济大学出版社,2021,111-139. * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111241154B (en) | Storage battery fault early warning method and system based on big data | |
| CN100465842C (en) | Data analysis system and method | |
| US20210011457A1 (en) | Online monitoring device and system for a 3D printing device | |
| CN104573845B (en) | Information system equipment state overhauling aid decision-making method | |
| US9734001B2 (en) | Efficient health management, diagnosis and prognosis of a machine | |
| US20160350671A1 (en) | Dynamically updated predictive modeling of systems and processes | |
| US20210157310A1 (en) | System and method for forecasting industrial machine failures | |
| CN102765643A (en) | Elevator fault diagnosis and early-warning method based on data drive | |
| CN114267178B (en) | Intelligent operation maintenance method and device for station | |
| CN101720453A (en) | System and method for predictive maintenance of a battery assembly using temporal signal processing | |
| CN114563150B (en) | Bridge health online detection module generation method, detection method, tool box and device | |
| CN102495320A (en) | Dynamic monitoring system and on-line monitoring method of transformer operation exception | |
| US20210150827A1 (en) | System and method for monitoring and predicting breakdowns in vehicles | |
| ES2963926T3 (en) | Method for monitoring the status of subsystems within a renewable generation plant or microgrid | |
| CN110703743A (en) | Equipment failure prediction and detection system and method | |
| CN116717437A (en) | Wind turbine generator system fault monitoring method and system | |
| CN113849774A (en) | Intelligent equipment maintenance platform based on life analysis and health monitoring | |
| KR20210050274A (en) | Equipment failure prediction system and method | |
| CN116403605B (en) | Stacker fault prediction method and related device | |
| CN117331790A (en) | Machine room fault detection method and device for data center | |
| CN116403605A (en) | Equipment fault prediction method, stacker fault prediction method and related devices | |
| JP7062505B2 (en) | Equipment management support system | |
| CN114492636B (en) | Transformer winding state signal acquisition system | |
| CN115687315A (en) | Fault diagnosis and health management system of fan | |
| CN109359421A (en) | A kind of real-time monitoring system and the lifting equipment based on real-time monitoring system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |